The idea that coffee is bad for your bones has been around for decades, and it is one of the more persistent pieces of nutrition folklore. The usual version goes like this: caffeine makes you lose calcium in your urine, calcium lost is calcium taken from your skeleton, and a lifetime of coffee therefore thins your bones and breaks your hip at eighty. Each step in that chain has been measured. Some of them hold up better than others, and the place the chain gives way is more interesting than either the scare or the reassurance.
What the worry is actually about
Bone is not inert. It is constantly being dismantled and rebuilt, and the calcium in your blood is drawn from your diet, your urine losses and, when those do not balance, from the skeleton itself. Bone mineral density is the standard measurement of how much mineral is packed into a given area of bone, and low density raises the risk of fracture, though it is not the only thing that does. Osteoporosis is the clinical version of that problem. So the question about coffee is really two questions: does caffeine measurably change calcium balance, and does that change ever show up as thinner bones or broken ones?
Caffeine does increase calcium in urine, briefly
The first link in the chain is real and has been measured directly. Kynast-Gales and Massey[1] kept 17 healthy volunteers in a metabolic ward on a controlled diet and gave them two doses of caffeine, at 3 mg per kilogram of lean body mass, at 7 and 10 in the morning. Urinary calcium and magnesium rose significantly for six hours after the second dose. Between 4pm and 1am there was no difference, and between 1 and 4 in the morning excretion actually fell below the caffeine-free day, which is the body attempting to conserve. The authors concluded that this night-time conservation was not enough to offset the morning loss, leaving a small net deficit over 24 hours.
That is the study most often quoted when coffee is blamed for osteoporosis, and it does show what it is said to show. What it does not show is how large the loss is in the context of a normal diet, or what happens to bone as a result.
How big is the calcium loss, really?
This is where the picture changes. Heaney[2], reviewing human physiological work and controlled calcium balance studies, concluded that caffeine has a clear but very small depressant effect on intestinal calcium absorption, and, across the balance studies, no effect on total 24-hour urinary calcium excretion. His estimate of the size of the effect is the memorable part: the negative effect of caffeine on calcium absorption is small enough to be fully offset by as little as one to two tablespoons of milk.
Heaney also proposed an explanation for why the population studies kept finding associations. People who drink a lot of caffeinated beverages tend to drink less milk, because the two compete for the same place in a diet. A high caffeine intake, in other words, is often a marker for a low calcium intake, and low calcium intake is a well-established cause of skeletal fragility. He noted that every observational study implicating caffeinated beverages had been done in populations consuming substantially less calcium than recommended, and found no evidence of harm to bone in people meeting the recommended intake.
What happens to bone density itself
If the calcium effect mattered at the skeleton, it would show up in bone density scans. Chen and colleagues[3] pooled 20 studies covering 508,312 people, splitting them at 2 cups a day for coffee. The pooled mean difference in bone mineral density for coffee was 0.020, with a confidence interval running from -0.003 to 0.044, which includes no difference. For hip fracture, the pooled hazard ratio was 1.008, with an interval from 0.760 to 1.337. Their conclusion was blunt: daily coffee or tea consumption was not associated with bone density or hip fracture risk.
Younger skeletons have been checked separately, because adolescence is when most bone mass is laid down. Cui and colleagues[4] combined a cross-sectional analysis of US national survey data with Mendelian randomisation, a method that uses gene variants affecting coffee intake as a natural experiment. Children and adolescents in the highest quarter of caffeine intake showed no significant difference in femur neck, total femur or total spine bone density compared with the lowest quarter, and the genetic analyses found no causal relationship either.
The fracture evidence, where coffee and caffeine part company
Fractures are what actually matter to a person, and the largest synthesis here found something genuinely odd. Asoudeh and colleagues[5] pooled 22 studies, 15 cohort and 7 case-control. Coffee consumption showed no significant association with fracture risk in either study type. Caffeine intake, analysed separately, did: a 15% higher risk overall, and in the dose-response analysis each additional 100 mg of caffeine a day was associated with about 2% greater risk, an association the authors themselves described as marginal.
In subgroup analysis the pattern split further. Higher coffee intake was associated with a 15% lower fracture risk in men. Studies with less than 12 years of follow-up found a 14% higher risk, while longer ones did not. When a finding flips direction depending on sex and study length, the most likely explanation is confounding rather than biology, and the authors' own summary, that coffee was protective in men while caffeine was positively associated with risk, is a description of an inconsistency rather than a resolution of one.
An earlier dose-response meta-analysis by Lee and colleagues[6] is worth reading alongside it, as is the Swedish cohort described next, which measured both fracture incidence and bone density in the same women rather than relying on one or the other.
The single most useful study here
If you read only one paper on this question, make it the Swedish Mammography Cohort analysis by Hallström and colleagues[7], because it measured both ends of the chain in the same population. They followed 61,433 women born between 1914 and 1948 from 1987 through 2008, with coffee intake assessed by repeated food frequency questionnaires rather than a single recall. Over that time 14,738 women had a fracture of any type and 3,871 had a hip fracture, and in a subgroup of 5,022 women bone density was measured directly.
The fracture results showed nothing. After multivariable adjustment, the hazard ratio per 200 ml of coffee was 0.99 for any fracture and 0.97 for hip fracture, both intervals hugging 1. The bone density results, though, showed something: women drinking 4 or more cups daily had 2% to 4% lower bone density than those drinking less than 1 cup, depending on the site measured. The odds ratio for actually having osteoporosis was 1.28, with an interval from 0.88 to 1.87, so not statistically clear.
That combination is the whole story of this field in one study. A high coffee intake was associated with a small reduction in bone density that did not translate into more broken bones. Density is a proxy, and a 2% to 4% difference is smaller than the differences that drive fracture risk in practice. It is also the kind of difference that could easily be produced by what heavy coffee drinkers eat, drink and do rather than by the coffee.
What about tea?
Tea is the obvious comparison, since it delivers caffeine too, and the same meta-analysis that cleared coffee looked at it. In Chen and colleagues[3], splitting at 1 cup a day, the pooled mean difference in bone density for tea was 0.039, with an interval from -0.012 to 0.09, and the pooled hazard ratio for hip fracture was 0.93, with an interval from 0.84 to 1.03. Both point gently in tea's favour and neither reaches significance.
That matters for interpreting the coffee findings. If caffeine were doing real damage to bone, the caffeinated drink consumed in the largest volumes worldwide would be expected to show it, and it does not. Tea also contains compounds coffee does not, so this is suggestive rather than decisive, but it fits the conclusion that caffeine is not the load-bearing variable here.
Why the studies disagree with each other
Three things make this literature messier than it looks. The first is the milk problem Heaney identified: coffee drinking and dairy intake are entangled, and a study that does not measure calcium intake well cannot separate them. The second is that caffeine comes from many sources, and a study measuring cups of coffee and a study measuring milligrams of caffeine are not measuring the same exposure, which is exactly where the fracture meta-analysis split. The third is smoking. Heavy coffee drinkers have historically been more likely to smoke, and smoking is a genuine risk factor for low bone density, so older studies that adjusted for it poorly will show coffee looking worse than it is.
The umbrella review by Poole and colleagues[8], which assessed meta-analyses across dozens of health outcomes, found that fracture risk in women was one of the few outcomes where higher coffee consumption looked unfavourable, against a general backdrop of benefit or no effect. It is a useful corrective in both directions: the harm signal is not imaginary, and it is also not large or consistent.
What this means if you drink coffee every day
On this evidence, the calcium that caffeine costs you is real but small, measured in a tablespoon or two of milk rather than in your skeleton. The bone density studies find nothing. The fracture studies find nothing for coffee, and something small and inconsistent for caffeine. Nobody has run a randomised trial giving people coffee for a decade and counting their fractures, and nobody is going to, so this will stay observational.
The practical implication is about calcium, not about coffee. If your calcium intake is adequate, the research gives no reason to cut coffee for your bones. If it is not, the research gives you a reason to fix the calcium. A flat white or a latte does both at once, which is one of the few times a nutrition question answers itself in a way that favours the nicer drink. For readers who take their coffee black and are worried, the relevant lever is dairy or a fortified alternative elsewhere in the day, not the coffee.
Where Filipino coffee culture fits
Milk is not a given in every coffee culture. Traditional kapeng barako in the Philippines is usually drunk black, often strong and often several times a day, while the café drinks that dominate in cities like Marikina, the lattes, flat whites and Spanish lattes, are milk-heavy. On the evidence above, the black-coffee habit is not a bone problem by itself, but it also does not supply the offsetting calcium, so the rest of the diet has to. For what goes into those milk-based drinks, our guide to milk alternatives covers which ones are fortified with calcium and which are not, since that difference matters more here than the coffee does.
The limits of this evidence
Almost everything above is observational. The controlled work, including the metabolic ward study, measured calcium excretion over hours and days rather than bone over years, and the participants were young and healthy. The meta-analyses pool studies that defined coffee intake differently, measured calcium intake unevenly, and followed people for very different lengths of time, which is visible in their heterogeneity statistics. Mendelian randomisation partly gets around confounding but relies on assumptions of its own. None of this establishes cause in either direction.
The practical answer
Coffee is not established as bad for your bones. Caffeine causes a small, short-lived rise in urinary calcium that a splash of milk covers, bone density studies find no meaningful difference, and the largest fracture analysis found no association for coffee at all, with a small and inconsistent one for caffeine. The one group the evidence suggests should pay attention is people with a low calcium intake, where coffee is less the problem than the diet it displaces. For related reading, see our article on whether coffee stunts growth, which covers the adolescent bone research in more detail, and our article on safe caffeine limits. Nothing here is medical advice, and anyone with diagnosed osteoporosis should take direction from their own clinician.